John Cutler · 12 min read · AI & Product
TBM 437: Tokens, Hours, Points, and Other Curious Proxies
Read the original on cutlefish.substack.com →
TBM 437 argues that the obsession with 'return on tokens' distorts value measurement in AI-enabled work, and that hours and capacity proxies often misrepresent true ROI. It shows how time allocation, context switching, and non-linear production dynamics complicate whether more time or more tokens actually yield more value. The piece advocates a broader, universal theory of value beyond easy-to-measure proxies and invites readers to reflect on how token-based ROI questions relate to real outcomes.
Key Takeaways
- Proxies like hours and tokens are easy to measure but can mislead ROI without considering quality, leverage, and the production function.
- Time is non-fungible and context switching and focus quality heavily influence outcomes; more hours don’t automatically equal more value.
- ROI in AI requires a broader theory of value and shouldn’t be driven solely by short-term, easily attributable use cases.
- The measurement problem is universal; AI adds twists but the foundational economics of value remain the same.
Excerpt
Everyone is talking about “return on tokens.” Vendors love it (as long as the news is good). Companies that had long since given up on any sort of value architecture or on understanding the ROI of product work are suddenly chomping at the bit. Which makes sense: they haven’t been able to lay off enough people to cover their token budgets while still hitting their promises to the board. It is hard to know what sits below the surface: a genuine effort to understand how AI might augment human capabilities, or some version of “how many people can we safely…